Estimating and Testing Linear Hypotheses on Parameters in the Log-Linear Model
Abstract
The minimum discrimination information theorem yields minimum discrimination information estimates as members of an exponential family which can be expressed in log-linear additive form. The author illustrates in terms of a particular contingency table, the relations implied by hypotheses and subhypotheses of interest on the values of the natural parameters and random variables of the exponential family, the related moment parameters, and the estimates of cell entries. Computations are carried out using the Deming- Stephan iterative algorithm and its extension by the generalized iterative scaling procedure of Darroch and Ratcliff.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 28, 1973
- Accession Number
- AD0763461
Entities
People
- S. Kullback
Organizations
- George Washington University